Protocols For High Efficiency Wireless Networks

F REQUENCY D IVISION M ULTIPLE A CCESSFDMAT IME D IVISION M ULTIPLE A CCESSTDMA 1.4.1 1.4.2 13 Basic considerations on the capacity of DS-CDMA systems CHAPTER 2: THE GLOBAL SYSTEM FOR MO

WIRELESS NETWORKS

WIRELESS NETWORKS

by

Alessandro Andreadis Giovanni Giambene

KLUWER ACADEMIC PUBLISHERS

NEW YORK, BOSTON, DORDRECHT, LONDON, MOSCOW

©2002 Kluwer Academic Publishers

New York, Boston, Dordrecht, London, Moscow

Print ©2003 Kluwer Academic Publishers

All rights reserved

No part of this eBook may be reproduced or transmitted in any form or by any means, electronic, mechanical, recording, or otherwise, without written consent from the Publisher

Created in the United States of America

Visit Kluwer Online at: http://kluweronline.com

and Kluwer's eBookstore at: http://ebooks.kluweronline.com

Dordrecht

The authors wish to thank Prof Giuliano Benelli for his continuous help and encouragement.

F REQUENCY D IVISION M ULTIPLE A CCESS(FDMA)

T IME D IVISION M ULTIPLE A CCESS(TDMA)

1.4.1

1.4.2

13

Basic considerations on the capacity of DS-CDMA systems

CHAPTER 2: THE GLOBAL SYSTEM FOR MOBILE COMMUNICATIONS

17

17

17 18

20 22 25 29 30 34 38

40 42

43

45

52 55 65 68

69 80 81

82 83 85

GSM-GPRS AIR INTERFACE: DETAILS ON PHYSICAL LAYER

EDGE AND E-GPRS

R ADIO RESOURCE MANAGEMENT CONCEPTS

Q O S ISSUES IN THE GPRS SYSTEM

UMTS TRAFFIC CLASSES

UMTS ARCHITECTURE DESCRIPTION

UTRA-FDD physical layer characteristics

Mapping of transport channels onto physical channels

UTRA-TDD physical layer characteristics

1

2 2 4 8

3.5

3.6

3.7

V OICE SERVICE IN UMTS

N EW SERVICE CONCEPTS SUPPORTED BY UMTS

UMTS RELEASES DIFFERENCES

R ESOURCE REUSE WITH TDMA AND FDMA

C ODE D IVISION M ULTIPLE A CCESS(CDMA)

85 86 87

91

93 93

101 102

103

104 106

107

115

120 123

127 135

135 136 139 143 146 147

151

151 152 153 154

155 156 156 157 157

CHAPTER 4:SATELLITE COMMUNICATIONS

4.1 B ASIC CONSIDERATIONS ON SATELLITE COMMUNICATIONS

4.1.1

4.1.2

4.1.3

Satellite orbit types

Frequency bands and signal attenuation

Satellite network telecommunication architectures

4.2 D IFFERENT TYPES OF MOBILE SATELLITE SYSTEMS

4.2.1

4.2.2

Satellite UMTS

Future satellite system protocols for high-capacity transmissions

4.3 O VERVIEW OF PROPOSED MOBILE SATELLITE SYSTEMS

CHAPTER 5:MOBILE COMMUNICATIONS BEYOND 3G

5.1

5.2

R EVIEW ON NEW ACCESS TECHNOLOGIES

4G VIEW FROM EU RESEARCH PROJECTS

PART II: SCHEDULING TECHNIQUES, ACCESS SCHEMES AND MOBILE INTERNET PROTOCOLS FOR WIRELESS COMMUNICATION SYSTEMS

CHAPTER 1: GENERAL CONCEPTS ON RADIO RESOURCE

S ELF- S IMILAR TRAFFIC SOURCES

D ATA TRAFFIC SOURCES

D ESCRIPTION OF LAYER 2 PROTOCOLS OF GPRS

M EDIUM ACCESS MODES

T ERMINAL STATES AND TRANSFER MODES

P-persistent access procedure

One- and two-phase access procedures

Queuing and polling procedures

Paging procedure

A detailed example of a one-phase access procedure

3.5 GPRS PERFORMANCE EVALUATION

CHAPTER 4: RRM IN WCDMA

169 170 172

175

176 177

180 183

183

187

188 190

192 192

193

196 200 201

205

207 211

217

217 219 219 220 223

R ADIO INTERFACE PROTOCOL ARCHITECTURE: DETAILS

T RANSPORT AND PHYSICAL CHANNELS

5.2.1

5.2.2

Spreading for downlink and uplink physical channels

Multiplexing, channel coding and interleaving

5.3

227

227 234 238 241

245 245 246 249

RLC SERVICES AND FUNCTIONS

R ESOURCE MANAGEMENT FOR DSCH

ATB-P PROTOCOL DESCRIPTION

ATB-P PERFORMANCE EVALUATION

T HE CLASSICAL PRMA PROTOCOL IN LEO-MSS S

PRMA WITH H INDERING S TATES(PRMA-HS)

M ODIFIED PRMA (MPRMA)

S TABILITY STUDY OF PACKET ACCESS SCHEMES

A NALYSIS OF R OUND R OBIN TRAFFIC SCHEDULING

2-MMPP TRAFFIC DELAY ANALYSIS

L ESSONS LEARNED ON RRM STRATEGIES

CHAPTER 9: A FIRST SOLUTION TOWARDS THE MOBILE INTERNET: THE WAP PROTOCOL

WAP PROTOCOL STACK

9.3.1 Bearers for WAP on the air interface

9.4 T OOLS AND APPLICATIONS FOR WAP

CHAPTER 10: THE MOBILE INTERNET

10.1 IP AND MOBILITY

258 259

263

264 265 266 266 267

269 283

Radio transmissions have opened new frontiers allowing the exchange

of information with remote units From the first applications oftelegraphy and radio broadcast, wireless transmissions have obtained agreat success with the widespread diffusion of mobile communications

We live in the communication era, where any kind of information must

be easy accessible to any user at any time Mobile communicationsystems are the technical support that allows the realization of suchconcepts

With the term mobile communications we embrace a set oftechnologies for radio transmissions, network protocols, mobileterminals and network elements

The widespread diffusion of wireless communications is makingnational borders irrelevant in the design, delivery and billing ofservices, thus requiring international coordination of standardizationefforts in order to evolve regional systems towards global ones

Parallel to the evolution of radio-mobile systems, we assist to themassive diffusion of Internet network and contents, thus allowing manyusers on the earth to be interconnected and to exchange any kind ofinformation, data, images and so on

Hence, there is a quick convergence of mobile communications and

Internet, i.e., mobile computing (see Fig 1).

The first cellular systems became operational at the beginning of 1980(first-generation, 1G) They employed analog techniques and rapidlydiffused with each country having its own system A first evolution wasachieved 10 years later by the adoption of digital standards (second-generation, 2G) Presently, we are assisting to the deployment of third-generation mobile cellular systems (3G) that under umbrellarecommendations collect at least three different standards They areintended to provide the users with high bit-rate transmissions so as toallow a fast access to the Internet and, in general, multimediatransmissions on the move [i],[ii].

In some European countries and in Japan the widespread diffusion ofmobile communications has reached the point to surpass the number ofwired phones This is an important achievement that significantlyhighlights the diffusion of mobile communication systems

The unique capabilities of new cellular systems are expected to provideusers with integrated multimedia applications Small, powerful,application-enabled devices will bring mobility needs together with thedesire for data and information Networks will be based on the IP

protocol [iii], including the support of Quality of Service (QoS) for

differentiated traffic classes

The air interface still represents the system bottleneck, by limiting theavailable user bit-rate due to both spectrum availability and radiopropagation impairments

At present, some mobile terminals have integrated a Java Virtual

Machine, an important step towards the mobile computing and the

support of typical Internet applications In fact, the Java language permitsthe development of platform-independent applications Another powerfultool for the realization of new applications and services is represented by

the eXtensible Markup Language (XML) and related technologies In

fact, XML can be used to design Web pages that can be adapted todifferent Internet access devices and technologies (e.g., mobile terminals

with small displays, Personal Digital Assistants, common personal computes, etc.) by using the characteristics of the HyperText Transfer

Protocol (HTTP) In fact, an Internet server can be equipped with an

adaptation engine that recognizes the access technology according tosuitable fields in the HTTP packet header; hence, different translationrules can be used to adapt the XML contents [iv]

However, the expected diffusion of new applications and multimediaservices can be only reached trough a novel system design that takesinto account all the communication aspects from the application layer

to the physical one, according to the OSI standard reference model.This approach is particularly effective for the air interface In fact, auser application cannot be designed without accounting for the limitedbandwidth, error resilience and reduced display sizes on mobileterminals In addition to this, the performance of the transport layerprotocol (TCP) must be evaluated in the presence of air interfaceresource constraints and the related traffic must be suitably managed toavoid that transmission delays or channel impairments negatively affectthe TCP throughput Moreover, the network layer must account for usermobility and the consequent re-routing of information when a userchanges its cell The frequency of handoff procedures among adjacent

cells will be exacerbated in future 3G micro-cellular systems Hence,the handoff process needs to be particularly optimized to avoid the loss

of information during handoffs Finally, the medium access controllayer must be able to integrate the support of different traffic classes,

guaranteeing ad hoc QoS levels, fairness among users and high

utilization of radio resources

All these aspects call for solutions suitably developed for the airinterface [v] Therefore, the focus of this book is on the optimization ofthe protocols at different layers in order to achieve simultaneously themaximum utilization of radio resources and the maximum satisfaction

of users, two aspects typically in contrast

This book will cover different wireless communication scenarios and,

in particular: 2.5G and 3G mobile communication systems (i.e., GPRS,UTRA-FDD and UTRA-TDD); 4G broadband wireless access systems(e.g., HIPERLAN/2); mobile satellite systems A complete review ofsuch systems is carried out in PART I Then, PART II will first focus

on both the performance evaluation of different resource managementtechniques for the above mentioned air interfaces and, then, willaddress the protocols at network and transport layers to allow themobile access to the Internet (i.e., TCP/IP and WAP) Hence, we willconsider the impact on the throughput of cellular systems due to boththe user mobility and the transmission of data packets on error-pronechannels

[i]

[ii]

[iii]

M Zeng, A Annamalai, V K Bhargava, “Recent Advances in

Cellular Wireless Communications”, IEEE Comm Mag., pp

128-138, September 1998

Ojanpera and R Prasad Wideband CDMA for Third Generation

Mobile Communications Artech House, October 1998.

T Robles, A Kadelka, H, Velayos, A Lappetelainen, A Kassler,

H Li, D Mandato, J Ojala, B Wegmann, “QoS Support for an

All-IP System Beyond 3G”, IEEE Comm Mag., pp 64-72,

August 2001

References

[v]

Network Working Group, “Hypertext Transfer Protocol HTTP/1.1”, (Web page) URL: http://ww.ieft.org/rcf/rcf2616.txt,June 1999

-M N Moustafa, I Habib, -M Naghshineh, -M Guizani,

“QoS-Enabled Broadband Mobile Access to Wireline Networks”, IEEE

Comm Mag., Vol 40, No 4, pp 50-56, April 2002.

In a wireless communication system, radio resources must be provided

in each cell to assure the interchange of data between the mobileterminal and the base station Uplink is from the mobile users to thebase station and downlink is from the base station to the mobile users.Each transmitting terminal employs different resources of the cell A

multiple access scheme is a method used to distinguish among different

simultaneous transmissions in a cell A radio resource can be a

different time interval, a frequency interval or a code with a suitablepower level All these characteristics (i.e., time, frequency, code andpower) univocally contribute to identify a radio resource [1] If thedifferent transmissions are differentiated only for the frequency band,

we have the Frequency Division Multiple Access (FDMA) Whereas, if

transmissions are distinguished on the basis of time, we consider the

Time Division Multiple Access (TDMA) Finally, if a different code is

adopted to separate simultaneous transmissions, we have the Code

Division Multiple Access (CDMA) However, resources can be also

differentiated by more than one of the above aspects Hence, hybridmultiple access schemes are possible (e.g., FDMA/TDMA)

In a cellular system, radio resources can be re-used between sufficiently

far cells, provided that the mutual interference level is at an acceptable

level This technique is adopted by FDMA and TDMA air interface,where the reuse is basically of carriers In the CDMA case, the number ofavailable codes is so high that the code reuse among cells (if adopted)does not increase the interference

In uplink, a suitable Medium Access Control (MAC) protocol is used to

regulate the access of different terminals to the resources of a cell that areprovided by a multiple access scheme [2] Whereas, in downlink the basestation has to transmit to the different users by means of a suitablemultiplexing scheme In the case of packet-switched traffics, there is also

a packet scheduling function that has to be implemented in the basestation

The frequency band available to the system is divided into differentportions, each of them used for a given channel (Fig 1); the differentchannels are distributed among cells (according to a reuse pattern).Adjacent bands have guard spaces in order to avoid inter-channelinterference First-generation terrestrial cellular systems (such as

Advanced Mobile Phone System, AMPS, that started operations in USA

on 1979) were based on analog transmissions with frequency modulationand FDMA [3] With the evolution towards digital communications, alsoTDMA and CDMA access schemes can be implemented

One disadvantage of FDMA is the lack of flexibility for the support ofvariable bit-rate transmissions, an essential prerequisite for futuremobile multimedia communication systems

In this scheme, each user has assigned the total bandwidth of a carrier for

transmission, but only for a short time interval (slot) that is periodically repeated according to a time-organization called frame.

Transmission is organized into frames, each of them containing a givennumber of slot intervals, to transmit packets of bits (Fig 2).

The classical multiple access techniques are described below [1]

1.1 Frequency Division Multiple Access (FDMA)

1.2 Time Division Multiple Access (TDMA)

For instance, let us refer to the transmission of speech through a digitalcommunication system The voice source signal (analogue signal) issampled with a suitable rate Each obtained value is then quantized with asuitable number of bits Then, a source coding scheme can be adopted toreduce the transmission bit-rate Finally, dynamic compression andpredictive schemes are adopted (accordingly, it is possible to achieve alow bit-rate voice transmission up to 2.4 kbit/s, for some satellitesystems) Thus, information bits are grouped in packets A voice sourcetypically require one packet to be transmitted a in a slot per frame (see thedarkest slots in Fig 2).

The US digital standard for cellular communications named IS-54 isbased on TDMA and tripled the capacity (= number of simultaneoususers supported per cell) with respect to the AMPS system, at a parity

of total bandwidth [3] The pan-European standard of generation cellular systems, GSM (Global System for Mobile

second-Communications), is based on TDMA More exactly, GSM adopts a

hybrid scheme of the FDMA/TDMA type: the available bandwidth isdivided among different 200 kHz sub-bands, each of them occupied by

a carrier accessed with a TDMA scheme

The main disadvantage of TDMA air interfaces is the high peaktransmit power that is required to send packets in the assigned slots.Moreover, a fine synchronization must be achieved at the beginning ofeach transmission for the alignment with the time-frame structure.Finally, a rigid resource allocation is supported by TDMA: according to

the above example of the voice traffic, one slot is assigned to a voicesource also during silent periods among talkspurts.

Cellular systems with TDMA or FDMA techniques are based on theresource reuse concept Indeed, due to the limited number of radioresources, it is necessary to reuse the same resource among sufficiently

distant cells so that the inter-cell interference is negligible The reuse

distance D, is the distance between two cells that may simultaneously

use the same channel (see Fig 3) Assuming a hexagonal regular

cellular layout for a given D value, it is possible to divide the total number of resources into K groups, distributed among the different cells as in a mosaic Possible values of K are: 1, 3, 4, 7, 9,

It is possible to prove the following relationship among D, R (the cell radius) and factor K [4]:

1.3 Resource reuse with TDMA and FDMA

For K = 7, we obtain the reuse mosaic shown in Fig 4 that corresponds

to C/I = 18 dB for

For the sake of completeness it is important to stress the fact that inpractical cases the cellular coverage is not hexagonal regular, butdepends on streets, building heights, obstacles, and so on A typicalexample can be the GSM 900 MHz cellular coverage shown in Fig 5,where different colors are related to different cells irradiated by three-sectored sites

The ratio between the power received at the base station from the

desired user of its cell, C, and the power received from co-channel users in cells at distance D, I, can be expressed as follows:

where is the path loss exponent (varying from 2 to 4, depending onthe cellular environment; is typical of free space propagation)

Let us refer to the classical circuit-switched voice service On the basis

of the reuse pattern K, if we have S system channels (i.e., frequency bands with FDMA or slots with TDMA), we may assign Q = S/K resources per cell (fixed channel allocation) Hence, at most Q

simultaneous circuit-switched phone calls can be managed per cell A

call generated in a cell where all its Q resources are busy is blocked and

cleared If we assume that calls arrive in a cell according to a Poisson

process with mean rate and that the channel holding time in a cell, X,

is generally distributed with mean value E[X], the blocking probability

experienced by a call is given by the well-known ERLANG-B

formula, according to an M/G/Q/Q model (M stands for Poisson arrivals; G means a general call duration time distribution; Q is the

number of requests in service = number of requests that can be hosted

by the system) [5]:

where Erlang.

The maximum cell capacity can be determined as the maximum load in

instance) If each user contributes an elementary load ofErlang, we may determine the maximum capacity of users per cell,

as:

Fig 6 shows the behavior of as a function of Q assuming that

each user contributes a load of 40 mErlang

The concept at the basis of CDMA spreading the transmitted signal over a

much wider band (“Spread Spectrum”, SS) Such techniques were

developed as jamming countermeasures for military applications in the

years 1950 Accordingly, the signal is spread over a band PG times

greater than the original one, by means of a suitable modulation based on

a PseudoNoise (PN) code1 [6]-[9]

PG is the so-called Processing Gain The higher PG, the higher the

spreading bandwidth and the greater the system capacity, as explainedlater in this Section Each user has its own code for uplink transmissions

In downlink, each base station has its code, but, in addition to this,suitable codes must be used to distinguish the different simultaneoustransmissions to the users in the cell

Even if a concentrated interfering signal is present in a portion of thebandwidth of the spread signal, the receiver de-spreads the useful signaland spreads on a wide band the interfering one, so that it becomes moresimilar to background noise

The receiver must use a synchronous code sequence with that of thereceived signal in order to correctly de-spread the desired signal

There are two different techniques to obtain spread spectrumtransmissions:

Direct Sequence (DS), where the user signal is multiplied by the PN

code with bits (named chips) whose length is basically PG times

smaller that that of the original bits This spreading scheme is well

suited for Phase Shift Keying (PSK) and Quadrature Phase Shift

Keying (QPSK) modulations (see Fig 7).

Frequency Hopping (FH), where the PN code is used to change the

frequency of the transmitted symbols (see Fig 8) We have fast

1

PN codes are cyclic codes that well approximate the random generation of 0 and 1 bits (e.g., Gold codes) These codes must have a high peak for the auto-correlation (synchronization purposes) and very low cross-correlation values (for the orthogonality

of different users).

1.4 Code Division Multiple Access (CDMA)

hopping if frequency is changed at each new symbol; whereas, a slowhopping pattern is obtained if frequency varies after a given number

of symbols The Frequency Shift Keying (FSK) modulation is well

suited for the FH scheme

A significant advantage of spread spectrum techniques for mobilecommunications is that they allow transmissions that are particularlyresistant to multipath fading (produced by reflections and diffraction ofthe signal due to the presence of obstacles along the signal path) In fact,spread transmissions by their nature mitigate the frequency selectiveaffects due to multipath fading [10]

The DS-CDMA technology is preferred to the FH-CDMA one, since it isexpensive to realize frequency synthesizers able to switch rapidly thetransmission frequency.

With DS-CDMA, a useful signal in a cell can be perfectly separated fromother DS-CDMA signals with different codes (interfering signals) in case

of synchronous transmissions with orthogonal codes (null correlation) If such synchronism is lost, partial cross-correlations among

cross-different codes loose the orthogonality, so that Multiple Access Interference (MAI) is experienced: the de-spreading process is unable to

conceal completely the interference coming from simultaneous users inthe cell This is the most common case in DS-CDMA cellular systems.Referring to uplink, MAI contributions come from simultaneoustransmissions in the same cell of the desired user and from adjacent cells.Note that synchronous transmissions can be naturally achieved fordownlink transmissions in a cell However, multipath phenomena maystill introduce some intra-cell MAI

Any technique able to reduce MAI increases capacity with DS-CDMA Inparticular, we may consider:

Squelching of transmissions during inactivity phases;

Multi-sector cells with directional antennas at the base station;

Multi-user receivers that reduce MAI coming from the users in thesame cell (intra-cell interference)

With CDMA transmissions, it is possible to use a special receiver, namedRAKE, that combines the signal contributions coming from differentpaths (micro-diversity) This receiver is particularly useful in themultipath environment of mobile communications in order to improve thebit error rate performance [10]

CDMA well supports powerful coding schemes that partly contribute tothe spreading process Accordingly, CDMA permits to achieve a greaterrobustness and a higher capacity than other multiple access schemes (i.e.,TDMA and FDMA) Hence, CDMA has been selected for future-generation mobile communication systems

CDMA needs that a power control scheme be adopted in order to avoidthat a user closer to the base station be received with an overwhelming

power with respect to users at cell borders (near-far problem) [8] Hence,

the signals of all the users must be received with the same power level(both for uplink and downlink), unless complex multi-user receivers areadopted

In general, multipath, shadowing and path loss phenomena call for acontinuous regulation of the transmitted power Channel propagation

variations are related to the Doppler frequency In Open-Loop Power Control (OLPC) schemes, the transmitter adapts the emission power depending on the power level of the received signal In Closed-Loop Power Control (CLPC) schemes the receiver notifies the received power

level to the sender that, thus, may vary the transmitted power to guarantee

an adequate received level OLPC and CLCP can be also adopted to

maintain a given received Signal-to-Noise and Interference Ratio (i.e.,

SNIR-based power control schemes)

For example, OLPC can be useful for quiescent mobile terminals that

continuously measure a beacon signal (i.e., the pilot signal) broadcast by

the base station in order to regulate the power they use for the first access.Such technique is adopted for random access transmissions (i.e., firstattempt after a silence phase) Whereas, CLPC can be adopted when acontinuous communication is established between a base station and amobile terminal

Third-generation cellular systems adopt a SNIR-based CLPC scheme

based on two different cooperating loops: the inner loop and the outer

one In particular, the outer loop continuously measures the signal qualityand defines the SNIR level value to be achieved (i.e., SNIR target) toguarantee a given error rate performance The inner loop continuouslyupdates the transmission power level so as to maintain the defined SNIRtarget value (one power control update measure is sent every 0.665 ms inthe WCDMA 3G system and every 1.25 ms in the cdma2000 3G system)

1.4.1 DS-CDMA spreading process

We consider a PSK modulation where antipodal rectangular bits (i.e.,+1 and -1) of duration are transmitted by multiplying them with a

carrier oscillation (see Fig 7) Before sending the signal, the bits aremodulated by a chip code sequence, where each chip has duration

This process is detailed in Fig 9, where we assume that the PNcode is a periodic sequence with period corresponding to the bitduration and that bits and chips have rectangular shapes (i.e., roll-offfactor equal to 0, for an ideal case)

1.4.2 Basic considerations on the capacity of DS-CDMA systems

We refer here to uplink and we assume that simultaneous transmissions

in the same cell and in adjacent cells contribute a MAI level that can be

modeled as Additive White Gaussian Noise (AWGN channel assumption) Let us consider having M equal mobile terminals perfectly

power-controlled in each cell of the DS-CDMA cellular systems (singletraffic case) We can write the following SNIR expression, assuming asingle-user correlation receiver and non-synchronous received signals

at the base station:

the impulse adopted to transmit a chip Hence, the spreading term

in (7) can be expressed as follows:

where:

P is the received power from a power-controlled user

is (approximately) the inter-cell to intra-cell interference ratio (depends on both channel characteristics and power control scheme;

a typical value is around 0.6, but even higher values are possible[11],[12])

W is the spreading bandwidth

is the single-sided background noise power spectral density

We may notice that (energy per bit multiplied by the bit-rate)

and we may consider I + N as the power of a white Gaussian noise over the bandwidth W with a given density Through some algebraicmanipulation, from (6) we obtain:

By substituting (8) in (7), we have:

We impose as Quality of Service (QoS) requirement:

is set to guarantee a given bit error rate value We assumethat the background noise level is negligible with respect to the useful

transmits according to an activity factor therefore, in previous

formulas, we may substitute to M Thus, (9) may be solved with respect to M in order to find the maximum capacity of simultaneous

users per cell:

CDMA systems adopt the soft-handoff scheme, so that while a userchange a cell, the signals received from (sent to) two cells arecombined in downlink (uplink) to improve the transmission quality bymeans of some degree of diversity Such technique improves thecommunication quality during a handoff, but it requires both a greaternumber of receivers (to avoid blocking events) and adequate powercontrol schemes We can roughly state that if 20% of cell area contains

users in soft-handoff, given M simultaneous users per cell, by reciprocity, the number of receivers at the base station must be M +

M/5 to support all the transmissions (including the duplicated ones due

to soft-handoffs)

The user capacity limit in DS-CDMA systems

increases with the processing gain and decreases

with both and the inter-cell interference factor

The base station must be provided with an adequate number ofreceivers in order to support the capacity expressed by (10) If a limitednumber of receivers is present, call blocking events may occur

Note that if a multi-user receiver is used at the base station, we mayneglect the intra-cell interference contribution in (6), so that the cellcapacity limit becomes:

This is also the typical condition for downlink transmissions, where wemay assume that orthogonality among users is preserved (with a goodapproximation, especially in micro-cellular systems2) Downlinkcapacity is not interference limited, but rather power limited, due to thefact that the base station has to divide its available power among all thesimultaneously transmitting users Hence, high power levels areneeded in the presence of far users, thus reducing the capacity ofsimultaneous users if equal power levels have to be guaranteed for allthe transmissions of a cell to avoid near-far problems with single-userreceivers

2

In macro-cellular systems, heavy multipath phenomena in urban areas may introduce intra-cell interference in downlink.

The GSM system originally defined in the 1980s was intended as aPan-European standard Today GSM has expanded into many parts ofthe World: GSM is widely adopted not only in Europe, but also inAustralia, Hong Kong, Singapore, South Africa and the UAE [13].

To provide additional capacity and to enable higher subscriberdensities, two other systems were added later: GSM 1800 (also known

as DCS1800) and GSM1900 (also named PCS 1900) Compared toGSM900, GSM 1800 and GSM 1900 differ primarily in the air interface.Besides adopting other frequency bands, they use a microcellularstructure (i.e., a smaller coverage radius for each cell) thus achieving acloser reuse of resources and, then, a higher capacity A very accuratehistory on the GSM standard can be found in [14] A complete survey

of other 2G systems is shown in [3]

Communications

2.1 Introduction to GSM

The GSM system is characterized by the building blocks that aredetailed below together with their interrelations, according to a typicalarchitecture that distinguishes between the base station sub-system, thenetwork sub-system and the mobile station (see Fig 11) [14]

2.1.1 Base station sub-system

Base Transceiver Station (BTS) is the base station with an antenna to

cover a cell (or more sectors)

The Base Station Controller (BSC): a group of BTSs is connected to a

particular BSC, which manages their radio resources The primary

function of the BSC is call maintenance Mobile Stations (MSs) send

reports of their received signal strengths to the BSC every 480 ms.With this information the BSC decides to initiate handovers to othercells, to change the BTS transmission power, etc

The Mobile Switching service Center (MSC) acts like a standard exchange of the Integrated Services Digital Network (ISDN) and

additionally provides all the functionality needed to support usermobility The main functions are: registration, authentication, locationupdating, handovers and call routing to a roaming subscriber If theMSC has also interconnections towards other networks, it is called

Gateway MSC (GMSC).

The Home Location Register (HLR) is a database used for the management of mobile subscribers It stores the International Mobile

Subscriber Identity (IMSI), Mobile Station ISDN Number (MSISDN)

and the address of the current Visitor Location Register (VLR) HLR

contains the information to route the call to the VLR where thedestination MS is currently registered The HLR also maintains the

description of the services associated with each mobile user (profile).

For each MS currently located in the geographical area controlled bythe VLR, the VLR contains the current MS location and selectedadministrative information from the HLR, necessary for call controland the provision of the subscribed services A VLR is connected toone MSC and is normally integrated into the MSC hardware

The Authentication Center (AuC) is a protected database that holds a

copy of each subscriber SIM card secret key that is used forauthentication and encryption over the radio channel AuC is normallylocated close to each HLR within a GSM network

The signaling between functional registers in the network sub-system

uses Signaling System 7 (SS7).

The Short Message Service Center (SMSC) enables subscribers to send and receive SMS (Short Message Service) messages in the cellular

network The SMSC temporarily stores SMS than cannot be delivereddue to an unreachable user (store-and-forward service) The SMSC islinked to an MSC, through which it sends and receives SMS TheSMSC obtain SMS routing information from the HLR

2.1.2 Network sub-system

The Equipment Identity Register (EIR) is a database that contains a list

of all valid mobile station equipment within the network, where each

mobile station is identified by its International Mobile Equipment

Identity (IMEI) EIR is formed by three databases:

The GSM core network is based on Mobile Application Part (MAP)

protocols [14]

The International Telecommunication Union (ITU), which manages

international allocation of radio spectrum (among many otherfunctions), has allocated the following bands:

White list for all known, good handsets

Black list for bad or stolen handsets

Grey list for handsets that are uncertain.

Uplink: 890 - 915 MHz (= mobile station to base station)

Downlink: 935 - 960 MHz (= base station to mobile station)

Voice service

SMS

Fax service

Data service

Transparent data (no error correction and constant delay)

Non-transparent data (error correction and variable delays)

Access to modems in telephony networks; i.e., access to the Internet

requires a circuit-switched call to an Internet Service Provider

A first evolution of the GSM standard was GSM-phase 2 (completed on1995), with the introduction of packet-oriented data transmission

services such as SMS and Unstructured Supplementary Service Data

(USSD):

An SMS contains at most 160 characters (140 octets) sent to/from

an MS via a signaling channel that depends on the status of the MS;SMS is a store-and-forward service provided by an SMSC:messages are kept within the SMSC until delivered to MS Paging

of the MS is needed to send each SMS message The bit-rate for thetransmission of an SMS is about 800 bit/s

Each USSD message contains at most 140 octets, conveyed on theair interface as SMS with bit-rate approximately of 800 bit/s USSD

is a transaction-oriented service with multiple mobile-originated ormobile-terminated messages during one session USSD traffic isalways exchanged with the HLR The GSM network can routeUSSD data to an external server

SMS can be used for applications like news, weather, stock exchange

or road traffic information Whereas, USSD are well suited fortransactional applications

The great success of SMS traffic highlights the need for packet datatransmissions on the move At present, 1 billion SMS messages are sentevery day It is expected that there will be 1.1 billions Internet users by

2004 and at least half of them will use wireless terminals to access the

Web A first answer to these needs is represented by the General

Packet Radio Service (GPRS) [15]-[20] that provides packet switched

services in an evolved GSM network (GSM-phase 2+) up to(theoretically, with no coding protection) about 170 kbit/s by assigning

8 slots of the same frame to a given user However, presenttechnological implementations allow up to 4 slots to be destined to thesame mobile terminal, thus achieving the potential maximum bit-rateper user of about 85 kbit/s

In order to reuse frequencies, GSM and GPRS adopt a combination ofFDMA and TDMA (i.e., reuse of carriers, each with TDMA resourcesharing scheme)

GPRS introduces new network nodes in the existing GSM system

architecture The most important ones are the Serving GPRS Support

Node (SGSN) and the Gateway GPRS Support Node (GGSN).

However, the GPRS architecture also introduces the followingelements:

Gb LLC (user data) and BSSGP (signaling) over Frame Relay (FR)

Gc Mobile Application Protocol (MAP); for location information

retrieval

Gd for short messaging over GPRS

Whereas, a simplified GPRS architecture is given in Fig 13, where thefollowing interfaces are shown (for the description of some relatedprotocols, see the following Section 2.8):

2.3 GPRS network architecture

Point-to-Multipoint Service Center (PTM-SC),

Border Gateway (BG),

Charging Gateway (CG),

Legal Interception Gateway (LIG).

A complete GSM-GPRS architecture is shown in Fig 12